Mobile computing devices, such as cellular telephones, utilize speech recognition technology to perform various voice-activated functions such as dialing and information queries. For example, a networked computing device may send a voice query to a voice portal platform for directory assistance, customer service, financial, travel, e-commerce, or theater information. The voice portal utilizes speech recognition techniques to process voice queries received over a network (such as the Internet) from the mobile computing device, determines an appropriate result, and returns the result back to the computing device. Current methods of utilizing network-based speech recognition however, suffer from a number of drawbacks including network latency and downtime. Thus, when latency in a network is high or the network is down, the roundtrip time for the receipt of results from a voice query is delayed or interrupted causing the user of a mobile computing device who made the query to be inconvenienced. It is with respect to these considerations and others that the various embodiments of the present invention have been made.